BEGIN:VCALENDAR
VERSION:2.0
PRODID:-//Talks.cam//talks.cam.ac.uk//
X-WR-CALNAME:Talks.cam
BEGIN:VEVENT
SUMMARY:Self-Healing Codes - Dr. Michael Rule\, University of Cambridge
DTSTART:20201105T140000Z
DTEND:20201105T150000Z
UID:TALK153790@talks.cam.ac.uk
CONTACT:Thiago Burghi
DESCRIPTION:Neural representations change over time\, even for habitual be
 haviors. This phenomena\, termed "representational drift"\, seems to be at
  odds with long-term stable neural representations. Previously\, we showed
  that representational drift was gradual\, and might be tracked using weak
  error feedback. In this talk\, I show how stable representations could be
  achieved without external error feedback. I'll discuss a model for repres
 entational drift\, which captures features of neural population codes obse
 rved experimentally: Tunings are typically stable\, but occasionally under
 go larger reconfigurations. I then discuss "self healing codes"\, which co
 mbine error-correction with neural plasticity. Self-healing codes can trac
 k drift without outside error feedback. The learning rule required is biol
 ogically plausible\, and amounts to a form of homeostatic Hebbian plastici
 ty. When combined with network interactions that allow neurons to share in
 formation\, such homeostatic plasticity could allow a subpopulation of sta
 ble cells to maintain an accurate readout of an unstable population code.
LOCATION:Online (Zoom)
END:VEVENT
END:VCALENDAR